From the /battlefield_comp Reddit, written by GottHoId
In the past weeks, we have supplied the E-sports Battlefield League with servers, so they could run a first test-cup. With each new release we gathered the data from these servers, to have a look, where the Meta of the game is going, how it currently plays and how well the mode does in regards of comebacks and such. We are obviously also gathering data from all matchmaker matches. But these are not as expressive in terms of the Meta as these ESB matches. The reason behind this is, a lot of people are just playing 1-3 rounds and after that, they stop playing. We can assume that players at that stage mostly try out Incursions and won’t know what they are doing. So, this generates a noise inside the Data, and prevents us from seeing any Meta. Furthermore, players that have already committed themselves to train and play with their teams deliver a way better data-set. Admittedly, it will still have a noise level, where lower skilled (this must be seen on the relative scale, on how good they already are) play the best teams in the cup (which played in previous titles ESL One, Pro series or in other top Divisions of the various leagues that exist in the community). Yet we assume the noise Level to be way lower, because teams always try to scrim equally matched opponents. Now let us get to the data and what we can learn from it.
In this graph, we can see how many games ended in a close result. During this test cup we changed on the matchmaker servers from first to 11 to first to 8. However we stuck first to 11 for this cup. 0 Set differences means it ended in a 10-10 and 11 Set differences means it ended in an 11-0. The Graph-labels which contain ‘old’ refer to the Incursions-Build 19th October to 1st November. And the ones that don’t refer to the build of the 2nd November to 16th November. As we can see, only 20% of the games ended close (11-7 => 10-10), while 51% of the games ended as a stomp (11-2 => 11-0). It took on average 3.84 Sets (479 seconds) for a team to start their comeback. A comeback in this case is, when a team manage to win their first set, after the first set of the game got won by the enemy team. The Cup started with 10 sets each round and normal experience gain speed, and so we maintained that in there too. This means whether the players already in the cup have a huge skill difference and they can’t find a matching opponent or our first game mode has a snowball effect. Therefor we are currently working on designs to push comeback moments.
In the Graph above you can see the top 15 selection of weapons, to which people died. We can observe that approximately 16% of all deaths are related to explosions and debris.
Here we can see how often each Kit got picked. As you might see there are 5 Kits, which get pretty much always picked, therefor we are trying to make other Kits more attractive and or nerf a bit other Kits.
Let us now find out, where the killing positions are at.
Here we can see a heat-map of Giant’s Shadow Redux. Spots that are hot (having a red to white tone) are positions where people tend to kill more than they die. Spots that are Cold (Having a green to blue tone) are positions where people die more than they kill. We can see here that roofs and high positions are good positions for killing people. Whereas the B house is almost a pure death trap. You can see that the entrances and windows have a red shimmer, which means usually people aggressively entering house or shooting from these spots tend to kill more, than they die. But staying and camping inside the B-house is definitely not a good idea. And here we can see why:
As you can see, the tank generates most of his kills on enemies that are in B-house or other spots where they have a wall behind them. This is due to the reason, that it is hard for the tank, to take out infantry before rank 2 (the tank unlocks a shotgun shell then). Further adding to this issue, that he has only 4 shots, and he sometimes needs to plan his shells through. So tankers rather wait for an infantry to have something next to them, so if they don’t hit its target directly that a close splash damage would kill it (if it is not an AT Assault). The bottom line of this is, don’t go into buildings if you have an enemy tank close by, try to make sure he has nothing to splash damage on, and make fast left and right movements, in the time he has a line of sight and nothing to splash on.
So let us take a look from where the tank does most of his kills and where he mostly dies at.
As you can see, the tank has 5 main spots, where he kills most from. While having at least on our first look no clear death position.
If we just look at where the tank dies the most,
We will see that there are actually hot spots, but comparing this to the heat-map above tells us, he makes way more kills from these positions either way.
Enough of the tank. We looked already at the power-positions of Giant’s Shadow Redux, but let us compare now, which power-position is more beneficial for which team-side.
The hot spots are here the power-spots for the left team, and the cold spots are the power-spots for the right team. Here we can see the left side has way more killing positions in B house and covers that are heading towards the right side back flag than vice versa.
If any of you have request for data or heat-maps for future posts, let me know! Also, if you enjoyed this one and want more, let me know. If you guys don’t want to see posts like these its fine as well, just tell me. We are always listening to feedback.
All Images can be found here aswell: https://imgur.com/a/dPTcP
Original Post: https://www.reddit.com/r/battlefield_comp/comments/7ezjiq/play_with_dice_gameplay_data_analysis/